From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Marco A Benatto <marco.antonio.780@gmail.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH] xfs_mdrestore: Don't rewind source file stream
Date: Sat, 3 Feb 2018 09:32:19 -0800 [thread overview]
Message-ID: <20180203173219.GE4849@magnolia> (raw)
In-Reply-To: <1517663796-89208-1-git-send-email-marco.antonio.780@gmail.com>
On Sat, Feb 03, 2018 at 11:16:36AM -0200, Marco A Benatto wrote:
> Today, xfs_mdrestore from stdin will fail if the -i flag is
> specified, because it attempts to rewind the stream after
> the initial read of the metablock. This fails, and
> results in an abort with "specified file is not a metadata
> dump."
>
> Read the metablock exactly once in main(), validate the magic,
> print informational flags if requested, and then pass it to
> perform_restore() which will then continue the restore process.
>
> Reported-by: Darrick J. Wong <darrick.wong@oracle.com>
> Signed-off-by: Marco A Benatto <marco.antonio.780@gmail.com>
> ---
> mdrestore/xfs_mdrestore.c | 55 +++++++++++++++++++++++------------------------
> 1 file changed, 27 insertions(+), 28 deletions(-)
>
> diff --git a/mdrestore/xfs_mdrestore.c b/mdrestore/xfs_mdrestore.c
> index 0bb4ac8..3db740d 100644
> --- a/mdrestore/xfs_mdrestore.c
> +++ b/mdrestore/xfs_mdrestore.c
> @@ -51,35 +51,34 @@ print_progress(const char *fmt, ...)
> progress_since_warning = 1;
> }
>
> +/*
> + * perform_restore() -- do the actual work to restore the metadump
> + *
> + * @src_f: A FILE pointer to the source metadump
> + * @dst_fd: the file descriptor for the target file
> + * @is_target_file: designates whether the target is a regular file
> + * @mbp: pointer to metadump's first xfs_metablock, read and verified by the caller
> + *
> + * src_f should be positioned just past a read the previously validated metablock
> + */
> static void
> perform_restore(
> FILE *src_f,
> int dst_fd,
> - int is_target_file)
> + int is_target_file,
> + const struct xfs_metablock *mbp)
> {
> - xfs_metablock_t *metablock; /* header + index + blocks */
> + struct xfs_metablock *metablock, tmb; /* header + index + blocks */
> __be64 *block_index;
> char *block_buffer;
> int block_size;
> int max_indices;
> int cur_index;
> int mb_count;
> - xfs_metablock_t tmb;
> xfs_sb_t sb;
> int64_t bytes_read;
>
> - /*
> - * read in first blocks (superblock 0), set "inprogress" flag for it,
> - * read in the rest of the file, and if complete, clear SB 0's
> - * "inprogress flag"
> - */
> -
> - if (fread(&tmb, sizeof(tmb), 1, src_f) != 1)
> - fatal("error reading from file: %s\n", strerror(errno));
> -
> - if (be32_to_cpu(tmb.mb_magic) != XFS_MD_MAGIC)
> - fatal("specified file is not a metadata dump\n");
> -
> + tmb = *mbp;
> block_size = 1 << tmb.mb_blocklog;
Why not avoid the structure copy and dereference the pointer directly?
block_size = 1 << mbp->mb_blocklog;
etc.
> max_indices = (block_size - sizeof(xfs_metablock_t)) / sizeof(__be64);
>
> @@ -211,6 +210,7 @@ main(
> int open_flags;
> struct stat statbuf;
> int is_target_file;
> + xfs_metablock_t mb;
struct xfs_metablock mb;
--D
>
> progname = basename(argv[0]);
>
> @@ -237,7 +237,12 @@ main(
> if (!show_info && argc - optind != 2)
> usage();
>
> - /* open source */
> + /*
> + * open source and test if this really is a dump. The first metadump block
> + * will be passed to perform_restore() which will continue to read the
> + * file from this point. This avoids rewind the stream, which causes
> + * restore to fail when source was being read from stdin.
> + */
> if (strcmp(argv[optind], "-") == 0) {
> src_f = stdin;
> if (isatty(fileno(stdin)))
> @@ -248,15 +253,12 @@ main(
> fatal("cannot open source dump file\n");
> }
>
> - if (show_info) {
> - xfs_metablock_t mb;
> -
> - if (fread(&mb, sizeof(mb), 1, src_f) != 1)
> - fatal("error reading from file: %s\n", strerror(errno));
> -
> - if (be32_to_cpu(mb.mb_magic) != XFS_MD_MAGIC)
> - fatal("specified file is not a metadata dump\n");
> + if (fread(&mb, sizeof(mb), 1, src_f) != 1)
> + fatal("error reading from file: %s\n", strerror(errno));
> + if (mb.mb_magic != cpu_to_be32(XFS_MD_MAGIC))
> + fatal("specified file is not a metadata dump\n");
>
> + if (show_info) {
> if (mb.mb_info & XFS_METADUMP_INFO_FLAGS) {
> printf("%s: %sobfuscated, %s log, %s metadata blocks\n",
> argv[optind],
> @@ -270,9 +272,6 @@ main(
>
> if (argc - optind == 1)
> exit(0);
> -
> - /* Go back to the beginning for the restore function */
> - fseek(src_f, 0L, SEEK_SET);
> }
>
> optind++;
> @@ -301,7 +300,7 @@ main(
> if (dst_fd < 0)
> fatal("couldn't open target \"%s\"\n", argv[optind]);
>
> - perform_restore(src_f, dst_fd, is_target_file);
> + perform_restore(src_f, dst_fd, is_target_file, &mb);
>
> close(dst_fd);
> if (src_f != stdin)
> --
> 1.8.3.1
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2018-02-03 17:32 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-02-02 19:11 [PATCH 1/2] xfs_mdrestore: Add -i option to built-in help Marco A Benatto
2018-02-02 19:11 ` [PATCH 2/2] xfs_mdrestore: Don't rewind source file stream Marco A Benatto
2018-02-02 20:27 ` Darrick J. Wong
2018-02-03 13:16 ` [PATCH] " Marco A Benatto
2018-02-03 13:18 ` Marco Benatto
2018-02-03 17:32 ` Darrick J. Wong [this message]
2018-02-05 11:50 ` Marco A Benatto
2018-02-05 16:47 ` Darrick J. Wong
2018-02-02 20:22 ` [PATCH 1/2] xfs_mdrestore: Add -i option to built-in help Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180203173219.GE4849@magnolia \
--to=darrick.wong@oracle.com \
--cc=linux-xfs@vger.kernel.org \
--cc=marco.antonio.780@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).